Grim Gjønnes, Sales & Marketing Ceetron, General Manager Crisp Ideas

Grim Gjønnes, Sales & Marketing Ceetron, General Manager Crisp Ideas

Some weeks ago, I came back from the NAFEMS* World Congress on CAE in San Diego where a question that I had thought resolved, was still (decades after the widespread adoption of 3D visualization technology) generating disagreement.

In preparing for my trip to NAFEMS, I had re-read a thread in LinkedIn’s CAE Professionals Group that explored (see the rationale for 3D visualization in CAE. The thread reached no explicit conclusions, but underneath it was a claim and an observation. The claim was that engineers make decisions based on numbers, not nice visual representations of engineering models. The observation (summed up nicely in the words of one participant), was that “Expensive 3D graphics adds zero value to the engineering decision[s] made [regarding] the product.”

I, like many others, begged to differ, and I will use today’s blog post to make this argument: Those who think 3D visualization has no or very limited strictly engineering role to play in bringing products to market, do not fully grasp the duality between visual analysis and numerical analysis.

Visualization is not just an effective way to support the core engineering processes of problem solving, collaboration, verification, and validation. In the new world of CAE—with its exponentially increasing complexity, collaboration, and model sizes—3D visualization is a necessary condition for the continuing evolution and utility of CAE. 3D visualization does not only create pretty pictures. It allows the human brain to synthesize vast amounts of data into a single compact representation of reality that allows for studying and navigating the details as well as the structure, across spatial scales, temporal scales, and relevant scalar / vector fields (say temperature, pressure, and stresses).

Going back to the NAFEMS World Congress, the event covered the future of CAE in quite some detail**. Among the topics were:

  • Private and public engineering applications and data in the cloud, and associated security concerns
  • New licensing/delivery models for engineering applications, including true SaaS
  • Multi-party, multi-site, multi-discipline collaboration
  • Large model sizes (driven by multi-scale (with regard to spatial scale), multi-level (for hierarchical assemblies of components), multi-physics, multi-body)
  • Appification
  • New materials
  • Immersive visualization
  • FEA and CFD used for biology modeling (ranging from DNA and proteins to Simulia’s Living Heart project)
  • The convergence between CAE and MBSE (Model-Based System Engineering)

To me, it seems obvious that these mega-trends will expand the role of 3D visualization in engineering decision-making. Three specific use cases illustrate why:

Engineering Analysis: A common model of engineering decision-making decomposes the process into:

  1. The problem statement, including requirements and constraints
  2. Specification of alternatives (what are they and how do they perform)
  3. The engineer’s preferences (or utility function)

In this process, the role of 3D visualization is profound, across engineering disciplines. It is by far the quickest and most effective tool for discerning spatial patterns, identifying hotspots, and formulating solution hypotheses, all of which are typically subsequently subjected to purely numerical analysis.

Collaboration: Today’s engineers work in collaborative groups, across different sites and organizational boundaries.  They need a mechanism that is a precise, single version of truth representation of reality and that is: easy to share; easy to manipulate; easy to communicate; and offers drill-down capabilities for extracting all relevant numerical information as needed.  This is even more so when needing a representation of reality that can be shared with non-specialists and across disciplines, often in the form of management, vendors, customers, and partners.  In-browser 3D visualization based on WebGL in combination with cloud-based modelling and model storage is precisely this mechanism; a table with numbers is not.

Verification & Validation: The CAE community puts significant emphasis on verification (getting the math right) and validation (getting the physics right), typically combined in a linear structure of activities.  If one looks at the various standards (e.g., ASME V&V 10, for FEA, ASME V&V 20, for CFD, and DNV-OS-F101, Appendix A, for FEA for submarine pipelines), one might conclude that V&V is all about verification of the computer code, and that the rest is just comparing computational results with test results, all done as a batch job. This is for a software and numerical modelling guy like me misleading; V&V is for companies with significant reuse of simulations models an iterative process that improves accuracy, robustness, and credibility over time, typically over years. In such extended V&V concept, personal experience from a range of fields (including underwater acoustics modelling, towed array modelling, and fracture mechanics for pipeline weld integrity assessments)  suggests that 3D visualization (together with reuse across projects) contribute to a high percentage of # issues identified and resolved as part of typical model V&V processes.  And, in Ceetron we systematically use visual comparison (in addition to automatic inspection of individual nodes or elements) when testing new releases of our 3D Components product.

Of course, the process of problem-solving is fundamentally the same across disciplines. It’s worth noting that McKinsey (a premier management consultancy whose regard for quantitative analysis is well-known), explicitly advises its consultants to use charts as a more effective way to communicate complex numerical relationships than tables***. It is also interesting to see how the big data community has started to see visualization as a tool complementary to purely numerical approaches, essentially for the same reasons as outlined in my bullet points above.  And, reverting to materials science, to quote a colleague of mine: “who would claim to analyze [microstructural phenomena] without observing them through a microscope?”

So, what are the implications for the CAE community:

  1. The new world of CAE will create a new set of requirements, including larger models, 3D visualization for cloud-based CAE, and distributed 3D visualization, the sharing of 3D models across geographical/organizational boundaries.
  2. 3D visualization must let CAE applications enable a rapid back-and-forth among different spatial scales, disciplines, and parts.
  3. 3D visualization has become integral to your engineering and decision-making workflows. It would help if organizations would recognize it as such, not as a point-solution to be bolted on a simulator for presentation purposes.
  4. 3D visualization platforms will develop into WebGL-based and cloud-based larger-footprint solutions / platforms, including collaboration, sharing, and workflows. Such anytime and everywhere platforms will support other parts of the engineering workflow, including specialized simulator offerings.  We will see industry dynamics for such platforms similar to those in other platform-dominated markets.
  5. Security will be key for such solutions. We will see the same focus on protecting and maintaining the integrity of stored data here that we have already seen in finance and other arenas.

All told, the NAFEMS event was a great conference for the CAE community, including academia, tool vendors, professional services firms, and end users alike.  We are looking forward to the next one in 2017.


*NAFEMS used to be the government-sponsored National Agency for Finite Element Methods and Standards in the UK. Today, it is the leading organizer of professional events for the CAE community (see for more info.)

**Thanks to Dennis A. Nagy from Beyond CAE who, in an invited presentation, gave an excellent introduction to the theme of the future of CAE. He kindly sent me his slide deck used, which I have used to identify some of the issues in the above bullets.

***Full disclosure: I used to work for McKinsey.